All of lore.kernel.org
 help / color / mirror / Atom feed
From: Peter Maydell <peter.maydell@linaro.org>
To: qemu-arm@nongnu.org, qemu-devel@nongnu.org
Cc: Richard Henderson <richard.henderson@linaro.org>
Subject: [PATCH v2 14/57] target/arm: Implement MVE VCLZ
Date: Mon, 14 Jun 2021 16:09:24 +0100	[thread overview]
Message-ID: <20210614151007.4545-15-peter.maydell@linaro.org> (raw)
In-Reply-To: <20210614151007.4545-1-peter.maydell@linaro.org>

Implement the MVE VCLZ insn (and the necessary machinery
for MVE 1-input vector ops).

Note that for non-load instructions predication is always performed
at a byte level granularity regardless of element size (R_ZLSJ),
and so the masking logic here differs from that used in the VLDR
and VSTR helpers.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
 target/arm/helper-mve.h    |  4 ++
 target/arm/mve.decode      |  8 ++++
 target/arm/mve_helper.c    | 97 ++++++++++++++++++++++++++++++++++++++
 target/arm/translate-mve.c | 38 +++++++++++++++
 4 files changed, 147 insertions(+)

diff --git a/target/arm/helper-mve.h b/target/arm/helper-mve.h
index e47d4164ae7..c5c1315b161 100644
--- a/target/arm/helper-mve.h
+++ b/target/arm/helper-mve.h
@@ -32,3 +32,7 @@ DEF_HELPER_FLAGS_3(mve_vldrh_uw, TCG_CALL_NO_WG, void, env, ptr, i32)
 DEF_HELPER_FLAGS_3(mve_vstrb_h, TCG_CALL_NO_WG, void, env, ptr, i32)
 DEF_HELPER_FLAGS_3(mve_vstrb_w, TCG_CALL_NO_WG, void, env, ptr, i32)
 DEF_HELPER_FLAGS_3(mve_vstrh_w, TCG_CALL_NO_WG, void, env, ptr, i32)
+
+DEF_HELPER_FLAGS_3(mve_vclzb, TCG_CALL_NO_WG, void, env, ptr, ptr)
+DEF_HELPER_FLAGS_3(mve_vclzh, TCG_CALL_NO_WG, void, env, ptr, ptr)
+DEF_HELPER_FLAGS_3(mve_vclzw, TCG_CALL_NO_WG, void, env, ptr, ptr)
diff --git a/target/arm/mve.decode b/target/arm/mve.decode
index 3bc5f034531..24999bf703e 100644
--- a/target/arm/mve.decode
+++ b/target/arm/mve.decode
@@ -20,13 +20,17 @@
 #
 
 %qd 22:1 13:3
+%qm 5:1 1:3
 
 &vldr_vstr rn qd imm p a w size l u
+&1op qd qm size
 
 @vldr_vstr ....... . . . . l:1 rn:4 ... ...... imm:7 &vldr_vstr qd=%qd u=0
 # Note that both Rn and Qd are 3 bits only (no D bit)
 @vldst_wn ... u:1 ... . . . . l:1 . rn:3 qd:3 . ... .. imm:7 &vldr_vstr
 
+@1op .... .... .... size:2 .. .... .... .... .... &1op qd=%qd qm=%qm
+
 # Vector loads and stores
 
 # Widening loads and narrowing stores:
@@ -61,3 +65,7 @@ VLDR_VSTR        1110110 1 a:1 . w:1 . .... ... 111101 .......   @vldr_vstr \
                  size=1 p=1
 VLDR_VSTR        1110110 1 a:1 . w:1 . .... ... 111110 .......   @vldr_vstr \
                  size=2 p=1
+
+# Vector miscellaneous
+
+VCLZ             1111 1111 1 . 11 .. 00 ... 0 0100 11 . 0 ... 0 @1op
diff --git a/target/arm/mve_helper.c b/target/arm/mve_helper.c
index 7282f56b18d..3d3b35d4848 100644
--- a/target/arm/mve_helper.c
+++ b/target/arm/mve_helper.c
@@ -181,3 +181,100 @@ DO_VSTR(vstrh_w, stw, 4, int32_t)
 
 #undef DO_VLDR
 #undef DO_VSTR
+
+/*
+ * The mergemask(D, R, M) macro performs the operation "*D = R" but
+ * storing only the bytes which correspond to 1 bits in M,
+ * leaving other bytes in *D unchanged. We use QEMU_GENERIC
+ * to select the correct implementation based on the type of D.
+ */
+
+static void mergemask_ub(uint8_t *d, uint8_t r, uint16_t mask)
+{
+    if (mask & 1) {
+        *d = r;
+    }
+}
+
+static void mergemask_sb(int8_t *d, int8_t r, uint16_t mask)
+{
+    mergemask_ub((uint8_t *)d, r, mask);
+}
+
+static void mergemask_uh(uint16_t *d, uint16_t r, uint16_t mask)
+{
+    uint16_t bmask = expand_pred_b_data[mask & 3];
+    *d = (*d & ~bmask) | (r & bmask);
+}
+
+static void mergemask_sh(int16_t *d, int16_t r, uint16_t mask)
+{
+    mergemask_uh((uint16_t *)d, r, mask);
+}
+
+static void mergemask_uw(uint32_t *d, uint32_t r, uint16_t mask)
+{
+    uint32_t bmask = expand_pred_b_data[mask & 0xf];
+    *d = (*d & ~bmask) | (r & bmask);
+}
+
+static void mergemask_sw(int32_t *d, int32_t r, uint16_t mask)
+{
+    mergemask_uw((uint32_t *)d, r, mask);
+}
+
+static void mergemask_uq(uint64_t *d, uint64_t r, uint16_t mask)
+{
+    uint64_t bmask = expand_pred_b_data[mask & 0xff];
+    *d = (*d & ~bmask) | (r & bmask);
+}
+
+static void mergemask_sq(int64_t *d, int64_t r, uint16_t mask)
+{
+    mergemask_uq((uint64_t *)d, r, mask);
+}
+
+/*
+ * mergemask() should never be passed an unknown type; catch this bug
+ * at compile time with a link error if we can, otherwise at runtime.
+ */
+#if defined(__OPTIMIZE__) && !defined(__SANITIZE_ADDRESS__)
+void unknown_mergemask_type(void *d, uint64_t r, uint16_t mask);
+#else
+static inline void unknown_mergemask_type(void *d, uint64_t r, uint16_t mask)
+{
+    abort();
+}
+#endif
+
+#define mergemask(D, R, M)                      \
+    QEMU_GENERIC(D,                             \
+                 (uint8_t *, mergemask_ub),     \
+                 (int8_t *,  mergemask_sb),     \
+                 (uint16_t *, mergemask_uh),    \
+                 (int16_t *,  mergemask_sh),    \
+                 (uint32_t *, mergemask_uw),    \
+                 (int32_t *,  mergemask_sw),    \
+                 (uint64_t *, mergemask_uq),    \
+                 (int64_t *,  mergemask_sq),    \
+                 unknown_mergemask_type)(D, R, M)
+
+#define DO_1OP(OP, ESIZE, TYPE, FN)                                     \
+    void HELPER(mve_##OP)(CPUARMState *env, void *vd, void *vm)         \
+    {                                                                   \
+        TYPE *d = vd, *m = vm;                                          \
+        uint16_t mask = mve_element_mask(env);                          \
+        unsigned e;                                                     \
+        unsigned const esize = sizeof(TYPE);                            \
+        for (e = 0; e < 16 / esize; e++, mask >>= esize) {              \
+            mergemask(&d[H##ESIZE(e)], FN(m[H##ESIZE(e)]), mask);       \
+        }                                                               \
+        mve_advance_vpt(env);                                           \
+    }
+
+#define DO_CLZ_B(N)   (clz32(N) - 24)
+#define DO_CLZ_H(N)   (clz32(N) - 16)
+
+DO_1OP(vclzb, 1, uint8_t, DO_CLZ_B)
+DO_1OP(vclzh, 2, uint16_t, DO_CLZ_H)
+DO_1OP(vclzw, 4, uint32_t, clz32)
diff --git a/target/arm/translate-mve.c b/target/arm/translate-mve.c
index 4cabdf7a69b..9eb6a68c976 100644
--- a/target/arm/translate-mve.c
+++ b/target/arm/translate-mve.c
@@ -29,6 +29,7 @@
 #include "decode-mve.c.inc"
 
 typedef void MVEGenLdStFn(TCGv_ptr, TCGv_ptr, TCGv_i32);
+typedef void MVEGenOneOpFn(TCGv_ptr, TCGv_ptr, TCGv_ptr);
 
 /* Return the offset of a Qn register (same semantics as aa32_vfp_qreg()) */
 static inline long mve_qreg_offset(unsigned reg)
@@ -160,3 +161,40 @@ static bool trans_VLDR_VSTR(DisasContext *s, arg_VLDR_VSTR *a)
 DO_VLDST_WIDE_NARROW(VLDSTB_H, vldrb_sh, vldrb_uh, vstrb_h)
 DO_VLDST_WIDE_NARROW(VLDSTB_W, vldrb_sw, vldrb_uw, vstrb_w)
 DO_VLDST_WIDE_NARROW(VLDSTH_W, vldrh_sw, vldrh_uw, vstrh_w)
+
+static bool do_1op(DisasContext *s, arg_1op *a, MVEGenOneOpFn fn)
+{
+    TCGv_ptr qd, qm;
+
+    if (!dc_isar_feature(aa32_mve, s) ||
+        !mve_check_qreg_bank(s, a->qd | a->qm) ||
+        !fn) {
+        return false;
+    }
+
+    if (!mve_eci_check(s) || !vfp_access_check(s)) {
+        return true;
+    }
+
+    qd = mve_qreg_ptr(a->qd);
+    qm = mve_qreg_ptr(a->qm);
+    fn(cpu_env, qd, qm);
+    tcg_temp_free_ptr(qd);
+    tcg_temp_free_ptr(qm);
+    mve_update_eci(s);
+    return true;
+}
+
+#define DO_1OP(INSN, FN)                                        \
+    static bool trans_##INSN(DisasContext *s, arg_1op *a)       \
+    {                                                           \
+        static MVEGenOneOpFn * const fns[] = {                  \
+            gen_helper_mve_##FN##b,                             \
+            gen_helper_mve_##FN##h,                             \
+            gen_helper_mve_##FN##w,                             \
+            NULL,                                               \
+        };                                                      \
+        return do_1op(s, a, fns[a->size]);                      \
+    }
+
+DO_1OP(VCLZ, vclz)
-- 
2.20.1



  parent reply	other threads:[~2021-06-14 15:24 UTC|newest]

Thread overview: 77+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-06-14 15:09 [PATCH v2 00/57] target/arm: First slice of MVE implementation Peter Maydell
2021-06-14 15:09 ` [PATCH v2 01/57] target/arm: Provide and use H8 and H1_8 macros Peter Maydell
2021-06-14 15:09 ` [PATCH v2 02/57] target/arm: Enable FPSCR.QC bit for MVE Peter Maydell
2021-06-14 15:09 ` [PATCH v2 03/57] target/arm: Handle VPR semantics in existing code Peter Maydell
2021-06-14 15:09 ` [PATCH v2 04/57] target/arm: Add handling for PSR.ECI/ICI Peter Maydell
2021-06-14 19:15   ` Richard Henderson
2021-06-14 15:09 ` [PATCH v2 05/57] target/arm: Let vfp_access_check() handle late NOCP checks Peter Maydell
2021-06-14 15:09 ` [PATCH v2 06/57] target/arm: Implement MVE LCTP Peter Maydell
2021-06-14 15:09 ` [PATCH v2 07/57] target/arm: Implement MVE WLSTP insn Peter Maydell
2021-06-14 19:20   ` Richard Henderson
2021-06-14 15:09 ` [PATCH v2 08/57] target/arm: Implement MVE DLSTP Peter Maydell
2021-06-14 15:09 ` [PATCH v2 09/57] target/arm: Implement MVE LETP insn Peter Maydell
2021-06-14 15:09 ` [PATCH v2 10/57] target/arm: Add framework for MVE decode Peter Maydell
2021-06-14 15:09 ` [PATCH v2 11/57] target/arm: Implement MVE VLDR/VSTR (non-widening forms) Peter Maydell
2021-06-14 19:29   ` Richard Henderson
2021-06-14 15:09 ` [PATCH v2 12/57] target/arm: Implement widening/narrowing MVE VLDR/VSTR insns Peter Maydell
2021-06-14 19:39   ` Richard Henderson
2021-06-14 15:09 ` [PATCH v2 13/57] target/arm: Move expand_pred_b() data to translate.c Peter Maydell
2021-06-14 19:41   ` Richard Henderson
2021-06-14 15:09 ` Peter Maydell [this message]
2021-06-14 20:00   ` [PATCH v2 14/57] target/arm: Implement MVE VCLZ Richard Henderson
2021-06-14 15:09 ` [PATCH v2 15/57] target/arm: Implement MVE VCLS Peter Maydell
2021-06-14 15:09 ` [PATCH v2 16/57] bitops.h: Provide hswap32(), hswap64(), wswap64() swapping operations Peter Maydell
2021-06-14 15:09 ` [PATCH v2 17/57] target/arm: Implement MVE VREV16, VREV32, VREV64 Peter Maydell
2021-06-14 20:03   ` Richard Henderson
2021-06-14 15:09 ` [PATCH v2 18/57] target/arm: Implement MVE VMVN (register) Peter Maydell
2021-06-14 15:09 ` [PATCH v2 19/57] target/arm: Implement MVE VABS Peter Maydell
2021-06-14 20:06   ` Richard Henderson
2021-06-14 15:09 ` [PATCH v2 20/57] target/arm: Implement MVE VNEG Peter Maydell
2021-06-14 15:09 ` [PATCH v2 21/57] tcg: Make gen_dup_i32() public Peter Maydell
2021-06-14 20:12   ` Richard Henderson
2021-06-14 15:09 ` [PATCH v2 22/57] target/arm: Implement MVE VDUP Peter Maydell
2021-06-14 20:15   ` Richard Henderson
2021-06-14 15:09 ` [PATCH v2 23/57] target/arm: Implement MVE VAND, VBIC, VORR, VORN, VEOR Peter Maydell
2021-06-14 15:09 ` [PATCH v2 24/57] target/arm: Implement MVE VADD, VSUB, VMUL Peter Maydell
2021-06-14 15:09 ` [PATCH v2 25/57] target/arm: Implement MVE VMULH Peter Maydell
2021-06-14 15:09 ` [PATCH v2 26/57] target/arm: Implement MVE VRMULH Peter Maydell
2021-06-14 15:09 ` [PATCH v2 27/57] target/arm: Implement MVE VMAX, VMIN Peter Maydell
2021-06-14 15:09 ` [PATCH v2 28/57] target/arm: Implement MVE VABD Peter Maydell
2021-06-14 15:09 ` [PATCH v2 29/57] target/arm: Implement MVE VHADD, VHSUB Peter Maydell
2021-06-14 15:09 ` [PATCH v2 30/57] target/arm: Implement MVE VMULL Peter Maydell
2021-06-14 15:09 ` [PATCH v2 31/57] target/arm: Implement MVE VMLALDAV Peter Maydell
2021-06-14 15:09 ` [PATCH v2 32/57] target/arm: Implement MVE VMLSLDAV Peter Maydell
2021-06-14 15:09 ` [PATCH v2 33/57] include/qemu/int128.h: Add function to create Int128 from int64_t Peter Maydell
2021-06-14 15:09 ` [PATCH v2 34/57] target/arm: Implement MVE VRMLALDAVH, VRMLSLDAVH Peter Maydell
2021-06-14 21:27   ` Richard Henderson
2021-06-14 15:09 ` [PATCH v2 35/57] target/arm: Implement MVE VADD (scalar) Peter Maydell
2021-06-14 21:30   ` Richard Henderson
2021-06-14 15:09 ` [PATCH v2 36/57] target/arm: Implement MVE VSUB, VMUL (scalar) Peter Maydell
2021-06-14 15:09 ` [PATCH v2 37/57] target/arm: Implement MVE VHADD, VHSUB (scalar) Peter Maydell
2021-06-14 15:09 ` [PATCH v2 38/57] target/arm: Implement MVE VBRSR Peter Maydell
2021-06-14 15:09 ` [PATCH v2 39/57] target/arm: Implement MVE VPST Peter Maydell
2021-06-14 15:09 ` [PATCH v2 40/57] target/arm: Implement MVE VQADD and VQSUB Peter Maydell
2021-06-14 15:09 ` [PATCH v2 41/57] target/arm: Implement MVE VQDMULH and VQRDMULH (scalar) Peter Maydell
2021-06-14 15:09 ` [PATCH v2 42/57] target/arm: Implement MVE VQDMULL scalar Peter Maydell
2021-06-14 15:09 ` [PATCH v2 43/57] target/arm: Implement MVE VQDMULH, VQRDMULH (vector) Peter Maydell
2021-06-14 15:09 ` [PATCH v2 44/57] target/arm: Implement MVE VQADD, VQSUB (vector) Peter Maydell
2021-06-14 15:09 ` [PATCH v2 45/57] target/arm: Implement MVE VQSHL (vector) Peter Maydell
2021-06-14 21:43   ` Richard Henderson
2021-06-14 15:09 ` [PATCH v2 46/57] target/arm: Implement MVE VQRSHL Peter Maydell
2021-06-14 15:09 ` [PATCH v2 47/57] target/arm: Implement MVE VSHL insn Peter Maydell
2021-06-14 15:09 ` [PATCH v2 48/57] target/arm: Implmement MVE VRSHL Peter Maydell
2021-06-14 15:09 ` [PATCH v2 49/57] target/arm: Implement MVE VQDMLADH and VQRDMLADH Peter Maydell
2021-06-14 15:10 ` [PATCH v2 50/57] target/arm: Implement MVE VQDMLSDH and VQRDMLSDH Peter Maydell
2021-06-14 15:10 ` [PATCH v2 51/57] target/arm: Implement MVE VQDMULL (vector) Peter Maydell
2021-06-14 15:10 ` [PATCH v2 52/57] target/arm: Implement MVE VRHADD Peter Maydell
2021-06-14 15:10 ` [PATCH v2 53/57] target/arm: Implement MVE VADC, VSBC Peter Maydell
2021-06-14 21:58   ` Richard Henderson
2021-06-14 15:10 ` [PATCH v2 54/57] target/arm: Implement MVE VCADD Peter Maydell
2021-06-14 15:10 ` [PATCH v2 55/57] target/arm: Implement MVE VHCADD Peter Maydell
2021-06-14 22:01   ` Richard Henderson
2021-06-14 15:10 ` [PATCH v2 56/57] target/arm: Implement MVE VADDV Peter Maydell
2021-06-14 15:10 ` [PATCH v2 57/57] target/arm: Make VMOV scalar <-> gpreg beatwise for MVE Peter Maydell
2021-06-14 22:20 ` [PATCH v2 00/57] target/arm: First slice of MVE implementation no-reply
2021-06-14 22:22 ` Richard Henderson
2021-06-21 16:37   ` Peter Maydell
2021-06-21 17:13     ` Richard Henderson

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210614151007.4545-15-peter.maydell@linaro.org \
    --to=peter.maydell@linaro.org \
    --cc=qemu-arm@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=richard.henderson@linaro.org \
    --subject='Re: [PATCH v2 14/57] target/arm: Implement MVE VCLZ' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.